1. What is/are the importance of knowing the different levels of measurement
(nominal, ordinal, interval, and scale)? (10 points) Talking about the importance of knowing the different levels of measurement which are the nominal, ordinal, interval, and scale, there are 3 reason why it is important, to begin with, comprehending the degree of estimation assists you in identifying how to analysis the information from that variable. When you discover that an action is apparent, you comprehend that the mathematical qualities are really short codes for the longer names. Furthermore, considering the amount of approximation can help you figure out what kind of factual study to conduct for the attributes you've previously allocated. You know you'll never be able to average the information values or perform a t-test on the results if you assume an activity is self-evident. In addition perceive that there is a pecking order inferred in the degree of estimation thought. Assumptions will be less prohibitive, and information examinations will be less delicate, at lower levels of estimation. The current level contains all of the traits of the one below it and adds a novel, new thing at each step up the pecking order. In general, it is preferable to have a higher significant level of estimation than a lower one. 2. Differentiate dependent variables from independent variables. (10 points) The difference between dependent variable and independent variable. Independent variable has a big impact on dependent variable for its reliability or dependency on independent variable, when it can be manipulated or change by experimenter. And dependent variable relies in independent variable for it is the result or outcome through independent variable. Ex. when using mobile phones the amount of hours being spent while using is the independent variable while the dependent variable is quality of sleep or numbers of hours of sleep. So using mobile phones more hours can affect our quality of sleeping or even can affect our health. The independent is the cause while dependent variable is the result or effect. 3. What are parametric and non-parametric test? (10 points) Tests that make assumptions about the parameters of the population distribution from which the sample is derived are known as parametric tests. The assumption that the population data is regularly distributed is frequently made. Because non-parametric tests are "distribution-free," they can be applied with non-Normal variables. 4. When is it appropriate to make use of a parametric test or a non-parametric test? (20 points) It is appropriate to make use of a parametric test when the mean quietly more accurate that represent the center of distribution of data and if the size of the sample is large enough. On the other hand, use non-parametric test if the median is more reliable or accurate who represent as the center of the distribution of data and even if there’s a large sample size. 5. Enumerate the various parametric test tools and its non-parametric tool equivalent. Also indicate the use/utility of each tool (ex. T-test is a type of inferential statistic used to determine if there is a significant difference between the means of two groups, which may be related in certain features) (50 points) Parametric Test o T-test - used to test the null assumption that the means of both groups are the same. o ANOVA - test whether there is a significant difference between the averages of two or more groups. Non-parametric Test o Wilcoxon signed rank test, one sample - approach determines if the sample could have been drawn from a population with a median value that is hypothesized. o Wilcoxon rank sum test - assumes that the data are at least on an ordinal scale, that the samples are independent and randomly chosen, and that the populations are roughly the same shape. And all data points are ranked in order, the rank sum of each sample is calculated, and the difference between the rank sums is compared. o Kruskal-Wallis Test - examines if the median values of three or more independent samples differ in any way. The data values are ranked in ascending order, and the rank sums are computed before the test statistic is computed. o Friedman Test - When the same parameter has been tested under multiple settings on the same subjects, this is an alternative to repeated measures ANOVAs. o Sign Test - solely uses + or - signs and ignores the data's real values. As a result, it's useful when measuring values is difficult. o Run Test - analyzes the number of runs in a sequence of observations to determine its randomness. A run is the appearance of one or more similar observations in a row. o Kolmogorov – Smirnov Test for Normality - created to determine whether two random samples were chosen from the same distribution. The KS test's null hypothesis is that two distributions are identical. The distance between the two empirical distributions, determined as the biggest absolute difference between their cumulative curves, is the KS test statistic. o Spearman Coefficient of Rank Correlation - determine the quality and direction of their connection. 6. Provide the assumptions for the use of each parametric test and its non-parametric equivalent. What statistical tools could help us determine these assumptions? ex. normality test – we have Kolmogorov smirnov or shapiro wilk test. (50 points) Process flow chart – non-parametric data has two factors which are comparison and relationship/prediction. Under comparison we have two – two techniques under it we have Wilcoxon signed-rank test and Friedman’s repeated measures. And under relationship is Kendall. Cause and effect diagram – identify the problem, work out the major factors involved (site, task, people, equipment, control), identify possible causes, and analyze your diagram. Pareto diagram – decide what categories you will use to group items, decide what measurement is appropriate, decide what period of time the pareto chat will cover, collect data, subtotal the measurements for each category, determine the appropriate scale or the measurements being collected, construct and label bars for each category. Check sheet – we have histogram, bar chart, and pareto chart as a result of using check sheet. Taguchi methods – planning, conducting and evaluating results of matrix experiments to determine the best levels of control factors. Designed experiments - altering one variable in order to see if changes in one variable lead to changes in another.